B. Muestra de Acción
4.2. Efectividad de la Práctica Pedagógica Reconstruida
Likeness was found to differ significantly with stereotype condition: F(l/44) = 8.9213 p = 0.0046, and with the time of presentation of stereotypic information (i.e., before or after presentation of the target image): F(l/44) ~ 6.1487 p = 0.0171. The interaction between condition and stereotype presentation time was not significant: F (1/44) = 1.8294 p = 0.1831 (Appendix 13, Table 1).
Next, a large-scale analysis of variance was conducted on the full data-set to indicate the significance of any main effects or interactions. No significant effect was found. For stereotype condition, F(l/44) = 2.2813 p = 0.1381. For time of stereotype presentation, F(l/44) = 0.0063 p = 0.9370. Hence, stereotypes did not have different effects with different presentation times. For the interaction between stereotype condition and presentation time, F(l/44) = 0.2429 p = 0.6245. Regenerated images were not found to be given significantly higher ratings on any one trait than any other, F (2/88) = 2.7052 p =
0.0724; neither was this affected by stereotype
condition: F(2/88) = 0.9237 p - 0.4008, or stereotype
presentation time: F(2/88) = 1.0128 p = 0.3674, (Appendix 14, Table 1). Therefore, the stereotype presented appeared to have no significant effect upon traits present in reconstructed body-images, and neither did stereotypic information appear to have different effects when presented at different times.
The present experiment constituted one of the first in the literature to use the whole-body as target stimuli: hence the magnitude of the expected effects was not known. Therefore, as in Experiment 2b, it was decided to investigate the effects of the stereotype conditions on trait judgments in separate, smaller analyses of variance in an attempt to isolate experimental effects from the noise that may be be present in the data. Since time of stereotype presentation had no independent effect upon ratings given, (an effect also supported in a separate analysis of variance, see Appendix 15, Table 1,) these data were combined for analysis.
No significant effect of stereotype condition was found for aggressiveness ratings: F (1/46) = 2.3581 p = 0.1315, or for honesty ratings: F(l/46) = 0.0354 p = 0.8516. However, bodies regenerated under the "assault" condition were rated as being significantly
larger than those regenerated under the "theft"
condition, in line with the experimental hypotheses: F(l/46) = 4.1053 p = 0.0486 (Appendix 16, Tables 1 - 3) .
Discussion
This experiment found that a piece of stereotypic information embedded in a two-hundred word biography can significantly affect the way in which a target body-shape is regenerated. When a target was said to have committed a crime stereotypically linked with aggressiveness, regenerated body-shapes were rated by independent judges as significantly larger than when the target was said to have committed a crime linked with dishonesty but not aggressiveness (theft). Given the significant correlation reported with Experiment 1 between perceived body-size and perceived aggressiveness, this is precisely the result expected.
Two results that are somewhat inconsistent with this finding were also achieved however. No actual difference in the size of the regenerated images was discovered. Hence, judges' interpretations of this variable appear to have been based on some factor other than the actual size of the images they were studying. It is tempting to argue that this factor might have been the perceived aggressiveness of the images: but if this is so, then it is perhaps odd that no significant effect was discovered for the aggressiveness trait itself. The correlation found between body-size and aggressiveness is admittedly quite small, even though it is highly significant (r = 0.1359 p < 0.001). Nevertheless, one would have expected the stereotypic information that the target committed an assault to have had a direct effect upon
aggressiveness per se, and not solely on a correlate
of aggressiveness. Conceivably, this effect may be
traceable to an artifact of the experimental materials
used. As mentioned in Chapter 4, the biographies
presented to the subjects in the present experiment may suggest that the cause of the crime committed by the target was not a dispositional one. Subjects may not therefore have felt that the target was a dispositionally aggressive person, and hence did not generate body-images which reflected stereotypic attributes of a dispositionally aggressive target. It should however be noted that there was a trend in the expected direction for the aggressiveness construct: mean rated aggressiveness in the "assault" condition being 5.425, against 4.875 in the "theft" condition, a difference of 0.55 on a nine-point scale. It is evident that only the "assault" condition had any effect upon judges' ratings of traits in the regenerated body-shapes. Where targets were said to have been accused of theft, no significant different in rated honesty was found (a mean of 5.025 in the "theft" condition against a mean of 4.983 in the "assault" condition). As mentioned above, it is by no
means immediately obvious just what the
stereotypically "honest/dishonest" body-shape might consist of; and, indeed, one aim of including this condition in the present experiment was to establish to what extent whole-body stereotypes parallel facial stereotypes, and whether there is indeed any such thing as a social stereotype of an "honest/dishonest"
body-type. The results of this experiment give cause to doubt whether any such stereotype does in fact exist. This is unsurprising when it is considered that the "theft" manipulation in Experiment 2a did not have a significant effect upon the rated honesty of regenerated faces, despite the fact that effects might be thought more likely to obtain when facial than whole-body stimuli are used.
' It could, however, be argued that another factor played the most important role in these results. It was found that the mean likeness of regenerated images to original stimuli was higher in the "theft" than in the "assault" condition. Conceivably, the main effect for body-size found here was caused by greater similarity in one condition, (or dissimilarity in the other,) to the original stimuli. Alternatively, a third variable may have had affected both body-size and similarity. It is conceivable that one of the two stereotypes presented in this experiment might have acted as a retrieval schema for the target information (see, e.g., Snyder and Uranowitz, 1978c): however this is a very tenuous conclusion to draw from the limited evidence available here. This possibility is discussed in far greater detail in Experiments 7 - 8 , below.
Experiment 3b
This experiment is a conceptual replication of Experiment 2b, utilising body rather than facial information for target stimuli. Two conditions were
run, in which subjects were told that the target individual received either a 2i or a 2ii degree. It is hypothesized that, where subjects are told a target received a 2i degree, then a more attractive body- shape should be generated than when subjects are told the target received a 2ii degree. Since this effect may be confounded by the effect of perceived
intelligence, (see the discussion of this possibility
in Chapter 3,) ratings of intelligence were also taken. It is hypothesized that body-shapes generated under the "2i" condition will be judged more intelligent than those regenerated under the "2ii" condition.
Methodology
The methodology for this experiment was identical with that of Experiment a, above, except that the biography did not mention that the target had been accused of any crime, but instead varied the class of degree said to have been awarded, i.e., either a 2i or a 2ii degree.
Subjects: A further forty-eight subjects were used
in this experiment. These were undergraduate and postgraduate student volunteers, who participated for a small fee. Twenty-nine subjects were female; nineteen were male.
Results
a') Actual Distortions of Regenerated Images: Mean
percentage actual distortion and standard deviations
are presented in Table 5.3, below.
Condition 2i I 2ii I ______ I Mean % I Distortion +26.835 I +26.459 I SD 10.093 I 10.766